Here are my tips for putting together a successful panel for an academic conference. Committees are happy to get panel applications rather than individual abstracts because it means less work for them. These are simple tips that have worked for me in the past. Your experiences may differ, of course, depending on your field of study and the particular conference, the vicissitudes of the conference committee, and the "trendiness" of the topics that year.
-Find another person whose work and your own have the same broader subject matter or approach. It helps if they are from a different institution or department from your own, but it's not necessary. Don't be afraid to venture outside your own level: professors and postdocs and grad students can happily be housed together. I have found it best when you are at least acquainted beforehand.
-Ask them to join your panel. If yes, you are now a twosome, and it gets easier. (If no, repeat step one.)
-Brainstorm for (an)other speaker(s) to complement your panel. This can be the most difficult part. Who will be able to round out the potentially narrow appeal of your twosome? Who will add a historical, analytic, or material dimension to the treatment of the topic? Who is available and hasn't already been scooped up by another panel? Don't be afraid to dream big or to approach people you haven't met in person yet. Often they will be flattered to be asked, and since there are already two of you, you will seem to have momentum. If they aren't available, don't be afraid to ask if they can recommend anyone else. Keep at this until you have the number of panelists you need.
-Think about each other's work and how it can come together in a way that makes sense as a whole. What areas of commonality are there? Are there broader questions that you're all addressing in your own way? If the panel application requires a description of the panel as a whole, this will be useful.
-Have each panel member give you an abstract, put them together, and submit your application before the deadline. Make sure that you follow the instructions in the CFP. This can include format issues, making sure to blind your submission, etc. Also make sure that the email gets sent (I know someone who missed out on a conference because his abstract got stuck in his email outbox...)
-Wait for the conference's decision. Good luck with this part.
-Once you are accepted, stay in touch with your panelists. There are often logistics to work out: who will chair the panel? How long will the talks be? Will your question period(s) come after each talk or at the end? Depending on the conference, you will have more or less opportunity to shape the session exactly the way you want to. It's also good to clarify whether everyone will bring a laptop or whether you'll use a shared one, whether you have A/V needs, etc.
-Try to arrange a face-to-face meeting at the conference before your session. This helps you recognize each other before the day of your session and bring each other up to speed on any changes to your papers.